85 results about "Adaptive coding" patented technology

The invention discloses a rolling bearing fault diagnosis method based on a multi-scale residual attention network and an adaptive Transform encoder. The rolling bearing fault diagnosis method comprises the following steps: acquiring original vibration data in the running process of a rolling bearing; segmenting the collected original vibration data into samples with specified lengths, and dividing the samples into a training data set and a test data set; inputting the training data set into a multi-scale residual attention network to perform preliminary multi-scale feature extraction; inputting the feature information extracted by the multi-scale residual attention network into an adaptive Transform encoder to obtain time sequence features; finally obtained feature information is subjected to GAP processing and then is input into a Softmax layer for fault diagnosis; the forward propagation calculation and the back propagation calculation are repeatedly executed to optimize model parameters until the diagnosis accuracy and loss of the training data set reach a stable level; and inputting the test data set into the trained model for fault diagnosis, and determining the health condition of the rolling bearing. According to the method, the adaptability and the diagnosis accuracy in time sequence dependence scenes such as rolling bearing fault diagnosis are enhanced.

Computer-implemented methods and systems for processing geometry replacements are disclosed. The methods include decoding / encoding a syntax element associated with a coding mode from / into a bitstream associated with geometry displacements; and reconstructing / converting, based on a coefficient configuration associated with the coding mode, a plurality of quantized transform coefficients from / to a plurality of zero-run length codes.

The invention relates to the technical field of data compression, in particular to an intelligent data compression method and system based on multi-protocol heterogeneous equipment interconnection. The method comprises the following steps: when it is detected that a new heterogeneous device applies to join a network, carrying out communication protocol real-time analysis and protocol conversion adaptation to obtain a heterogeneous device adaptation protocol; carrying out multi-device data stream receiving according to a heterogeneous device adaptation protocol, and generating data streams of a plurality of time windows; adaptive coding and multi-level compression are carried out on the data streams of the multiple time windows, and a multi-device adaptive compression mechanism is constructed; identifying compression load characteristics of each heterogeneous device, and performing dynamic task allocation to obtain an intelligent task allocation strategy; and driving real-time data intelligent compression operation based on an intelligent task allocation strategy and a multi-device adaptive compression mechanism. According to the invention, efficient and adaptive data compression of different devices is realized, and the data transmission efficiency and quality are improved.

The invention relates to the technical field of audio and video emergency transmission, and discloses an audio and video low-delay return method and system in an extreme environment. The method comprises the following steps: acquiring original multi-modal data of audio and video acquisition equipment in a target area, and analyzing a data state of the original multi-modal data; and meanwhile, available network transmission links are monitored, and the quality is evaluated. Self-adaptive coding parameters are generated in combination with the link quality and the data state, dynamic coding is executed on original data, and a coding stream suitable for redundant transmission is formed. And carrying out cooperative distribution transmission based on the network link set and the coded stream to generate return data. According to the generation process, a device and energy management instruction is formed and issued to the acquisition and relay device. According to the method, the network state and the content characteristics are optimized in a dynamic coding link in a collaborative manner, and the audio and video quality, the real-time performance and the system energy efficiency which are transmitted back in an extreme environment are improved through a feedback closed loop for transmitting a result to equipment management.

The invention discloses a coding and tagging method and system based on urban infrastructures, and belongs to the technical field of unified identification code coding, the realization of the method comprises the following steps: constructing a multi-level coding structure which comprises administrative division codes, facility type codes, unique serial numbers and dynamic attribute extension segments; acquiring real-time state data of the facility through an Internet of Things sensor, and writing key parameters into the dynamic attribute extension section; selecting an adaptive code identification mode according to a facility physical environment to carry out identification deployment; binding the codes with the three-dimensional coordinates of the urban information model to realize digital twinborn mapping; and fault prediction and intelligent decision making are realized through a coding depth integration and AI analysis module. According to the method, the problems of fragmentation, static identification information lag and data islanding of a traditional coding system are solved, the facility positioning efficiency is improved, the operation and maintenance cost is reduced, and reliable support is provided for intelligent management of the whole life cycle of urban infrastructures.

The invention belongs to the field of star flash mouse communication, and discloses a star flash mouse low-power-consumption communication system and method based on Polar code and SLE collaborative optimization, the system adopts a layered architecture design, collaborative optimization of energy efficiency and performance is realized through deep coupling of a physical layer and a link layer, and the performance of the system is improved. The core of the system is composed of three functional units including a Polar code coding module, an SLE control engine and a cross-layer optimization interface, and a complete closed-loop control system is formed. According to the invention, through deep cooperation of the Polar code and the SLE mechanism, the overall power consumption is reduced by 42.7% compared with the traditional scheme; according to the Polar code dynamic construction method based on track prediction, an ultra-low bit error rate of 10 <-5 > magnitude can still be kept in a 2.4 GHz complex interference environment, and in combination with signal-to-noise ratio self-adaptive coding parameter adjustment, it is ensured that transmission delay within 0.5 ms is always kept from a low-speed office scene to a high-speed electronic sports scene; in the aspect of engineering implementation, the method has the capability of quickly adapting to hardware platforms of different manufacturers, and a complete solution with high performance and low cost is provided for large-scale commercial use of the satellite flash technology in the consumer electronics field.

The invention relates to the technical field of medical image segmentation, in particular to a high-efficiency deep supervised distillation method and device applied to medical image segmentation, and the device comprises a pre-training coding module, a high-efficiency multi-scale fusion block, a multi-scale grouping attention gate, a high-efficiency frequency domain boundary extraction block, a high-efficiency fusion up-sampling block, and knowledge alignment multi-scale distillation. According to the invention, while the multi-scale features are effectively captured and fused, the feature expression capability in jump connection is enhanced; besides, the boundary prediction capability of the model is improved by using frequency domain information, and the multi-scale knowledge of the teacher model and the student model is gradually aligned in the encoding-decoding stage, so that the lightweight student model can still keep excellent segmentation performance under the condition of greatly reducing the calculation cost, and the method has extremely high compatibility, and is suitable for popularization and application. The method can be adaptive to multi-layer visual encoders of different scales, can be adaptive to output of different layers of the encoders, and effectively reduces the calculation amount and the parameter amount while guaranteeing the high segmentation performance.

The invention relates to the field of personnel re-identification, and provides an infrared-visible light personnel re-identification method and system based on wavelet representation, a wave adaptive encoder is designed, characteristics are decomposed into different frequency sub-bands through discrete wavelet transform, high-frequency and low-frequency information is processed through a high-low wave sensing unit respectively, and the personnel re-identification accuracy is improved. Adaptively adjusting the importance of different frequency band characteristics by using a wave frequency domain attention mechanism; a frequency enhancement module is introduced, modeling is carried out on global spectral characteristics by using Fourier transform, deep fusion of frequency domain and spatial domain information is realized through Einstein matrix multiplication, and the robustness of the model to different illumination conditions and imaging quality is enhanced; according to the method, similarity distribution clustering loss is enhanced, cross-modal features of the same identity are drawn close, features of different identities in the same modal are also explicitly pushed away, a more robust feature space is constructed, the problem of infrared image homogeneity is effectively solved, and the discrimination capability of the feature space is enhanced.

The invention designs an unmanned aerial vehicle video coding and decoding model based on deep learning and a processing method thereof. The unmanned aerial vehicle video coding and decoding model comprises an image compression reference module based on deep learning, a multi-dimensional self-adaptive coding and decoding module based on an attention mechanism and a video supplement module based on diffusion. The image compression reference module is used for extracting a high-fidelity and low-redundancy image potential representation and restoring the image potential representation into a reconstructed video frame; the multi-dimensional adaptive coding and decoding module is used for fusing the image frame dimension and the image space dimension to obtain a multi-dimensional feature map; and the video supplement module is used for generating more complete and accurate feature representation as video supplement information, so that the quality and integrity of the whole video are improved.

The invention discloses an XBT-T5 observation data deviation analysis and temperature correction method, and relates to the technical field of ocean observation, and the method comprises the steps: obtaining the original temperature observation data of a target sea area based on an XBT-T5 front-end temperature sensor, calculating the depth, generating a temperature-depth profile, and constructing a temperature-depth database in combination with historical depth observation archive data; the method comprises the following steps: synchronously acquiring temperature and depth data by using a CTD temperature-salinity-depth instrument, and carrying out association pairing with a temperature-depth database to generate an XBT-T5 observation data deviation field under the same spatial-temporal scale; based on generation of an XBT-T5 observation data deviation field under the same spatial-temporal scale, designing a dual-channel adaptive encoder, and constructing a temperature-depth collaborative correction model; and according to the temperature-depth collaborative correction model, coupling the temperature-depth database and generating an XBT-T5 observation data deviation field under the same spatial-temporal scale, thereby realizing XBT-T5 observation data deviation analysis and temperature correction. The method improves the quality and application value of observation data.

The invention discloses an ultra-high-definition video stream adaptive coding method based on deep learning visual saliency, and the method comprises the steps: carrying out the five-scale Gaussian filtering processing and image pyramid construction of a video frame, and combining Sobel gradient, Laplacian edge and local binary pattern feature extraction to generate a multi-scale feature map; a pre-training saliency detection network is adopted, and a smooth saliency thermodynamic diagram is generated through processing of a feature adaptation layer, a residual encoder, a self-attention mechanism and a transposed convolution decoder; dividing the video frame into a high region, a middle region and a low region according to the saliency thermodynamic diagram, and establishing a regionalization coding parameter table; performing differentiated prediction modes, motion estimation and quantization strategies on different salient regions; and organizing coded data according to an H.265 / HEVC standard, and embedding the saliency thermodynamic diagram into supplementary enhancement information for transmission. According to the method, the important region concerned by the user can be intelligently identified, a differentiated coding strategy based on content semantics is realized, and the coding efficiency is remarkably improved.

The invention discloses a full-view digital section pathological cell detection method based on PAC-YOLO. The method comprises the following steps: S1, constructing a PAC-YOLO model, and embedding a block-by-block local space attention mechanism in a YOLOv5 framework to perform feature extraction; s2, introducing a self-adaptive encoder, and optimizing the spatial feature extraction capability; s3, processing feature maps of different scales by using a multi-scale feature fusion module; s4, data preprocessing is carried out, and model parameter initialization is completed; s5, calculating loss and updating the model weight by adopting a self-adaptive optimization algorithm; s6, the detection precision is evaluated through the verification set, and the model structure is adjusted; and S7, evaluating the applicability and robustness of the model by using the test set. The problems of insufficient pathological section cell detection precision and low real-time response performance in the prior art are effectively solved, and the detection precision and applicability of an automatic pathological diagnosis system are remarkably improved.

The invention discloses a communication edge cloud segmentation reasoning method, device and equipment for a large language model, a product and a medium. The method comprises the following steps: segmenting the large language model into a client model and a server model based on a preset division standard; reasoning data to be processed through a client model, compressing an intermediate result through an adaptive encoder, and deploying the client model and the adaptive encoder on an edge device; the compressed intermediate result is reconstructed by using an adaptive decoder, the reconstructed intermediate result is reasoned through a server model to obtain a new token, and the adaptive decoder and the server model are deployed in a cloud server; and obtaining new to-be-processed data based on the new token, and repeating the previous steps until a preset condition is met so as to obtain a final prediction result. According to the method, the communication overhead can be reduced, the data privacy protection can be enhanced, and the method is suitable for different network environments and computing resources and has relatively good flexibility and expandability.

The invention discloses a method for searching individuals in a distributed manner in a big data set, and aims to solve the technical problems of conflict avoidance and state uniqueness distribution in a large-scale node structure data set. The method comprises the following steps of undirected graph establishment, structure mapping, four-class state initialization, adjacency detection and syn-position adjustment until the states of all adjacent nodes are different. Compared with the prior art, the method is based on graph structure mapping, odd-even state classification and syn-position adjustment, is used for carrying out distributed search and accurate positioning on individuals in a big data set, constructs a globally unique and locally adaptive coding method system through four-color state division and local adjustment, and has the advantages of unique state traceability, accurate recheck, high efficiency and the like. The method is flexible in state compression and expansion, can be expanded into various states, is high in distributed deployment capability, is high in universality, and is suitable for multiple fields of plane maps, image modeling, graph neural networks and file systems.

The present invention relates to the technical field of artificial intelligence, and relates in particular to a method and an apparatus for training a human body recognition model, a device, and a medium. The method comprises: training a human body recognition model by using second input human body images and image labels of the second input human body images, so as to obtain a trained human body recognition model, wherein the trained human body recognition model is used for performing feature extraction on a human body image, and the human body recognition model comprises a preset adaptive encoder, and an image encoder in a trained image-text model. In the present application, the human body recognition model, which comprises the preset adaptive encoder and the image encoder in the trained image-text model, is trained, and during training, the preset adaptive encoder and the image encoder in the trained image-text model are used to separately perform feature extraction, so that features extracted by the adaptive encoder are fused into image features, enhancing feature discriminability and thereby improving the training accuracy of the human body recognition model.

The invention discloses a real-time video processing method and system based on artificial intelligence, and belongs to the technical field of computers. Through deep fusion of multi-source information and a regional adaptive coding mechanism, the problem of collaboration between content generation and network resource allocation in a high-interactivity live broadcast scene is solved. In the feature extraction link, video, audio and text streams are synchronously processed, so that three-dimensional perception of a live broadcast scene is realized; in the multi-mode driving instruction set generation process, audio event features and text interaction features are subjected to semantic association, so that content generation and real-time interaction are highly synchronized, and response delay caused by information splitting in a traditional scheme is avoided; according to the regional differentiation coding processing, resource allocation is dynamically adjusted according to the regional importance of visual semantic feature recognition and the network state, and the coding quality of a digital human region is preferentially guaranteed, so that the stability of core visual elements is maintained under the network fluctuation condition, and the intelligent level of real-time video processing and the user experience are improved.

The invention discloses a big data matching recommendation system and method based on a technology transfer platform, and relates to the technical field of artificial intelligence, and the system comprises a data processing module, a knowledge graph construction module, an intention analysis module, a similarity calculation module and a recommendation sorting module. Wherein the data processing module adopts a field adaptive encoder to realize heterogeneous data standardization; the knowledge graph construction module extracts technical entities and relationships based on a pre-training language model; the intention analysis module extracts core features of a supply party and a demand party through a double-decoder mechanism; the similarity calculation module fuses graph structure analysis and differentiation weight distribution and realizes dynamic compensation; and the recommendation sorting module generates a personalized recommendation result in combination with the market potential evaluation. According to the invention, the intelligent matching precision of the technology transfer platform is obviously improved, the precise docking of cross-field technology supply and demand is realized, and the method has important practical value.

The invention relates to an underwater image enhancement method based on a regional differentiation fusion network, and belongs to the technical field of image information processing. The method comprises the following steps: inputting an original underwater image, and executing self-adaptive histogram equalization processing for limiting contrast to generate a corresponding self-adaptive enhanced image; respectively inputting the original underwater image and the self-adaptive enhanced image into an original encoder and a self-adaptive encoder, and generating initial original features and self-adaptive enhanced features through a shallow feature extraction module; the original features and the self-adaptive enhancement features are input into a cross fusion Swin Transform module, and regional differentiation feature fusion is carried out; inputting the fused regional differentiation features into a reconstruction decoder, carrying out gradual decoding and reconstruction through interpolation up-sampling and convolution activation operation, and outputting an enhanced underwater image; in the training stage, a joint loss function is adopted to supervise and optimize an enhancement result, and the joint loss function comprises Charbonier loss and VGG loss.

The invention provides a dual-mode while-drilling azimuth resistivity measurement method and device and related equipment, and relates to the technical field of charging management. According to the technical scheme provided by the invention, the gradient mutation event in the resistivity curve is identified through the gradient feature, and the key geological interface and the abnormal region are accurately positioned. The data transmission mode is determined by collecting the drilling working condition parameters corresponding to the gradient sudden change event, so that the transmission strategy is matched with the actual drilling state. And the transmission priority score of each data is calculated based on the drilling working condition parameters, the gradient sudden change event and the resistivity curve, so that quantitative evaluation of data importance and timeliness is realized. And comparing the transmission priority score with a preset percentage threshold, and then distinguishing first data and second data to ensure that important data is preferentially transmitted and secondary data is delayed to be stored. The first data are transmitted after being adaptively coded, and the second data are stored in a layered manner, so that the requirements of real-time performance and integrity are considered under the limited transmission bandwidth.

The invention relates to the technical field of rarefaction, in particular to a tensor calculation architecture and method based on a transposed structured sparse mode, and the tensor calculation architecture comprises a storage format conversion unit which is used for converting a preset sparse matrix into a transposed sparse matrix stored in a two-dimensional compression format; and the adaptive coding and decoding unit is used for determining a plurality of input elements in the transposable sparse matrix through format conversion, merging the plurality of input elements to obtain a plurality of calculation load matrixes, and sending the plurality of calculation load matrixes to the target processing unit array for tensor calculation. Therefore, the problems that the existing sparse mode only considers the protocol dimension to cause significant accuracy loss, and the existing storage format introduces discontinuous and redundant memory access in the flexible structured sparse mode to cause low bandwidth utilization rate are solved.

This invention provides a lightweight governance and coordinated response method for urban rail transit video, comprising: collecting urban rail transit video streams and corresponding business data; establishing a unified spatiotemporal benchmark and data mapping through preprocessing; extracting static background images through background modeling; decoupling the video in real time into static background and dynamic foreground; and constructing a ring-shaped buffer at the edge; extracting structured semantic features from the dynamic foreground to form a lightweight video stream for transmission and storage; integrating business data and feature analysis; triggering a first-level wake-up command when the warning rules are met; extracting the original foreground pixel stream T seconds before the warning from the buffer; generating key event frames using adaptive ROI encoding; matching and recording coordinated response rules based on the warning results; and finally saving the lightweight video stream, key event frames, warning results, and response records.

This invention discloses a 3D seismic fault segmentation method based on 3D adaptive coding and cross-domain collaborative learning, which addresses the problems of scarce annotations, the inability of 2D models to effectively represent 3D structures, and unstable predictions across different work areas. The method first preprocesses the source domain labeled and the target domain unlabeled seismic data. A lightweight 3D adapter is inserted based on the frozen backbone parameters of the 2D visual model, and a dual-path encoder with spatial and depth perspectives is constructed, fusing features from both perspectives to form a joint 3D representation. Then, a lightweight 3D decoder is used to progressively restore resolution and generate a fault probability volume. The source domain uses binary cross-entropy and Dice loss for supervised learning; the target domain uses a heterogeneous bi-sub-network composed of a visual branch and a 3D convolutional branch. Pseudo-labels are adaptively generated through voxel confidence, and unsupervised training is achieved by combining consistency and difference regularization, thus enabling stable and reliable 3D fault segmentation across different work areas even without target domain labels.

The application discloses a full-media full-signal integrated management platform, and belongs to the technical field of information processing and communication. The application comprises: a timing perception intelligent core, a network transmission control module, and a full-media signal processing module. The timing perception intelligent core comprises a global clock synchronization module, a delay prediction compensation module, an intelligent signal routing module, and an adaptive coding and decoding engine. The network transmission control module comprises a network adaptive transmission engine, a multi-path load balancer, and a fault switching recovery module. The full-media signal processing module is constructed based on a micro-service architecture and comprises a multi-modal signal acquisition module, a signal preprocessing module, a media content management center, a timing synchronization processing tool, and a unified API gateway. The intelligent resource scheduling pool provides intelligent scheduling management of computing, storage, network, and AI computing power, and realizes resource demand prediction, multi-target optimization distribution, and automatic elastic expansion and contraction.

The invention relates to the technical field of emergency communication and audio and video transmission, and discloses an audio and video emergency return method and system based on cooperation of multi-source heterogeneous equipment. The method includes receiving a set of raw perceptual data from a plurality of independent devices, including uncompressed audio and video streams, sensor readings, and network probe data packets. And performing multi-link survivability evaluation according to the network detection data packet, constructing a transmission channel in real time, and generating a multi-modal network convergence transmission instruction. Based on the instruction and an original audio and video stream, content-aware analysis is performed to dynamically derive coding parameters, and an adaptive coding control signal is generated. And generating a multi-device collaborative management instruction for coordinating a device protocol stack and resource allocation in combination with the device state report and the network situation. And integrating the instructions and the information to form a return execution decision, and scheduling multiple devices to cooperatively complete self-adaptive return of the audio and video data. According to the invention, the link robustness and the content transmission efficiency of audio and video return in the emergency scene are improved.

In various examples, properties may be determined for image regions, where the image regions are indicated by output data generated using MLMs. An encoder may use the properties to generate encoded images using encoding quality settings for the image regions. When an encoded image is decoded and applied to the MLMs, corresponding output data may indicate an image region which is likely to correspond to an encoded image region of the encoded image, and which may be applied to at least one MLM. Thus, the properties for encoding an image region to an encoded image can be adapted to control the visual quality of an image region determined from a decoded version of the encoded image. The properties may be determined based at least on performance metric values for the MLMs or based at least on a ranking of the image regions.

This invention discloses a multi-terminal network audio broadcasting system, belonging to the field of audio broadcasting technology. It includes a central server and multiple distributed audio terminals. The central server comprises: a geographic information module for storing digital elevation models and the deployment locations of each audio terminal; an adaptive coding module for dynamically adjusting the redundancy coding rate of the audio stream based on link quality feedback reported by each audio terminal; and a routing calculation module for generating a multi-hop broadcast tree based on a terrain-compensated propagation loss model. Each audio terminal includes a positioning module for collecting its real-time geographic coordinates and reporting them to the central server. This invention achieves accurate path loss calculation, adaptive redundancy coding, dynamic power adjustment, intelligent broadcast tree construction, efficient packet loss recovery, and dynamic relay coverage, improving the coverage, stability, and energy efficiency of audio broadcasting.

The invention specifically relates to an unmanned cluster identification method based on motion topology reasoning, and belongs to the technical field of artificial intelligence and unmanned systems. The method comprises the following steps: firstly, collecting a historical track sequence of unmanned cluster nodes, and constructing an interpretable feature system containing three types of physical features of motion, interaction and topology; inputting the feature vector into a self-adaptive encoder containing dynamic graph construction, multi-head attention and time sequence aggregation components, and outputting a high-order cluster feature vector of time sequence aggregation; then, target identity recognition and node interaction topology reasoning are completed in sequence, and a cluster node future trajectory prediction result is output through a dual decoder; and finally, identifying and predicting performance through joint optimization of a multi-objective loss function, and updating model parameters in real time to adapt to a dynamic task scene. According to the method, the problems of poor scale adaptability, task decoupling and feature black box of an existing method are effectively solved, and end-to-end joint optimization of identity recognition and trajectory prediction is realized.

The invention discloses a structure-first image compression method, which comprises the following steps of: acquiring an image to be compressed, and performing topological feature extraction to obtain topological feature data representing image connectivity and a hole structure; structural key point detection is carried out, and a structural key point set influencing the topological stability of the image is identified; combining to generate a topological importance thermodynamic diagram; and carrying out adaptive coding on the to-be-compressed image to generate compressed image data keeping the integrity of the key structure. According to the method, the encoder can be guided to skip fine quantization of a non-key area, and intelligent redistribution of computing resources is realized.

The invention relates to a computer-implemented method (100) for compressing image data, comprising at least the following steps: receiving (102) at least one image signal indicating at least one image content (20); providing (104) the image signal as at least one first input signal to at least one trained, adaptive encoder algorithm (22), which, from the first input signal, provides at least one first output signal indicating at least one coding (28, 30) that is uniquely assigned to the image content (20); providing (106) the image signal as at least one second input signal to at least one trained, adaptive text-generating algorithm (24), which, from the second input signal, provides at least one second output signal indicating at least one text (32) that describes the image content (20); providing (108) at least one transmission signal (34) indicating the coding (28, 30) and the text (32).The method (100) has increased efficiency.

The invention relates to an adaptive coding compression method and system for a column database, belongs to the technical field of data compression, and solves the problems of low compression rate and low query performance caused by curing of a compression algorithm and strong manual dependence in the prior art. Comprising the following steps: partitioning collected original data, and extracting a multi-dimensional feature vector from each dynamic column in each data block; inputting the multi-dimensional feature vector of each dynamic column into a trained prediction model to obtain a prediction result, the prediction result comprising a compression algorithm selection probability, a compression ratio prediction value and a decompression overhead level; based on the prediction result of each dynamic column, selecting an optimal compression algorithm for the dynamic columns through a multi-objective decision-making method; and further executing a compression operation. And the compression speed and precision of mass data are improved.